Evaluating Impact of Student Debt on Early Career Choices

美国科学界和工业团体观察到，美国没有产生足够数量的熟练本土工人来满足需求，特别是在推动创新的物理科学、计算机科学和工程领域。以成绩为基础的援助，例如奖学金和奖学金，既降低了学生投资高等教育的直接成本，也降低了学生毕业后累积高额学生债务的长期风险。本项目将评估优等生奖对女性和少数族裔学生在STEM学位课程中以及毕业后在科学队伍中不断增加的保留效果和后果。该项目考察了任何观察到的择优录取的影响程度，即未被充分代表的学生的早期职业选择，特别是由于降低了学生累积债务的风险。这项研究旨在为政策和制度战略提供信息，以留住美国学生和毕业生从事科学和工程职业。此外，由于科学和技术领域的毕业生也有更高的预期终身收入，鼓励历史上未被充分代表的种族/少数民族、第一代大学生和其他经济上处于不利地位的个人更广泛地参与这些领域，可能有助于解决持续存在和日益严重的社会经济不平等问题。该项目将采用线性和非线性计量经济学估计因果推理，利用两个现有的数据源。首先，利用一家大型公共少数族裔服务机构新开发的机构管理数据，并将学生水平的观察结果与公开可用的统计数据相匹配，该项目将评估收入条件下的奖学金，特别是美国教育部SMART助学金，或之前的社区大学或副学士学位转学分，是否会影响对STEM感兴趣的本科学生完成STEM学位。然后，利用美国国家科学基金会国家科学与工程统计中心的限制使用许可下的纵向调查数据，该项目将评估不同高等教育资助机制对学生毕业时累积债务负担的影响，并进一步评估较高的累积学生债务是否会阻碍STEM领域的应届毕业生从事理工科职业。

The U.S. scientific community and industry groups have observed that the United States is not producingsufficient numbers of skilled native workers to meet demand, especially in the physical sciences,computer science, and engineering fields that drive innovation. Merit-based aid, for example,scholarships and fellowships, reduces both the immediate cost to students of investing in highereducation, and students' longer-term risk of high cumulative student debt loads after graduation. Thisproject will evaluate both efficacy and consequences of merit awards for increasingretention of women and racial/ethnic minority students in STEM degree programs, and in the scientificworkforce after graduation. The project examines the extent of any observed effects of merit aidon underrepresented students' early career choices are specifically due to reducing students' risk of highercumulative debt. This research aims to inform policies and institutional strategies to retain U.S. studentsand graduates in science and engineering careers. In addition, because graduates in scientific andtechnical fields also have greater expected lifetime earnings, encouraging broader participation in thesefields by historically underrepresented racial/ethnic minorities, first-generation college students, and othereconomically-disadvantaged individuals may help to address persistent and growing socioeconomicinequality.The project will employ linear and nonlinear econometric estimation for causal inference, leveraging twoexisting data sources. First, using newly-developed institutional administrative data from a largepublic Minority-Serving Institution with student-level observations matched to publicly availablestatistical data, the project will assess whether income-contingent merit aid in particular, the U.S.Department of Education SMART grants, or prior community college or associate's degree transfercredit affected STEM degree completion among observably STEM-interested undergraduate students.Then, using longitudinal survey data under restricted-use license from the NSF National Center forScience and Engineering Statistics, the project will estimate the influence of different higher educationfunding mechanisms on students' cumulative debt load at graduation, and further whether highercumulative student debt deters recent graduates in STEM fields from pursuing science and engineeringoccupations.